The NIKA2 large-ﬁeld-of-view millimetre continuum camera ...NIKA2 and the IRAM 30-meter telescope represent today ideal tools to address these scientiﬁc questions, and many oth-1

Astronomy & Astrophysics manuscript no. NIKA2˙submitted˙rev3 c© ESO 2017November 28, 2017

The NIKA2 large-field-of-view millimetre continuum camera for the30m IRAM telescope

R. Adam1, A. Adane2, P.A.R. Ade3, P. Andre4, A. Andrianasolo5, H. Aussel4, A. Beelen6, A. Benoıt7, A. Bideaud7,N. Billot8, O. Bourrion9, A. Bracco4, M. Calvo7, A. Catalano9, G. Coiffard10, B. Comis9, M. De Petris11, F.-X. Desert5,

S. Doyle3, E.F.C. Driessen10, R. Evans3, J. Goupy7, C. Kramer8, G. Lagache6, S. Leclercq10, J.-P. Leggeri7,J.-F. Lestrade12, J.F. Macıas-Perez9, P. Mauskopf13, F. Mayet9, A. Maury4, A. Monfardini7?, S. Navarro8, E. Pascale3,

L. Perotto9, G. Pisano3, N. Ponthieu5, V. Reveret4, A. Rigby3, A. Ritacco8, C. Romero10, H. Roussel14, F. Ruppin9,K. Schuster10, A. Sievers8, S. Triqueneaux7, C. Tucker3, and R. Zylka10

1 Laboratoire Lagrange, Universite et Observatoire de la Cote d’Azur, CNRS, Blvd de l’Observatoire, F-06304 Nice, France2 University of Sciences and Technology Houari Boumediene (U.S.T.H.B.), BP 32 El Alia, Bab Ezzouar 16111, Algiers, Algeria3 Astronomy Instrumentation Group, University of Cardiff, The Parade, CF24 3AA, United Kindgom4 Laboratoire AIM, CEA/IRFU, CNRS/INSU, Universite Paris Diderot, CEA-Saclay, 91191 Gif-Sur-Yvette, France5 Univ. Grenoble Alpes, CNRS, IPAG, 38000 Grenoble, France6 Aix Marseille Universite, CNRS, LAM (Laboratoire d’Astrophysique de Marseille), F-13388 Marseille, France7 Institut Neel, CNRS and Universite Grenoble Alpes (UGA), 25 av. des Martyrs, F-38042 Grenoble, France8 Institut de RadioAstronomie Millimetrique (IRAM), Granada, Spain9 Laboratoire de Physique Subatomique et de Cosmologie, Universite Grenoble Alpes, CNRS, 53, av. des Martyrs, Grenoble, France

10 Institut de RadioAstronomie Millimetrique (IRAM), Grenoble, France11 Dipartimento di Fisica, Universita di Roma La Sapienza, Piazzale Aldo Moro 5, I-00185 Roma, Italy12 LERMA, CNRS, Observatoire de Paris, 61 avenue de l’Observatoire, Paris, France13 School of Earth and Space Exploration and Department of Physics, Arizona State University, Tempe, AZ 8528714 Institut d’Astrophysique de Paris, CNRS (UMR7095), 98 bis boulevard Arago, F-75014 Paris, France

Received December XX, XXXX; accepted XXX XX, XXXX

ABSTRACT

Context. Millimetre-wave continuum astronomy is today an indispensable tool for both general astrophysics studies (e.g. star for-mation, nearby galaxies) and cosmology (e.g. CMB - cosmic microwave background and high-redshift galaxies). General purpose,large-field-of-view instruments are needed to map the sky at intermediate angular scales not accessible by the high-resolution inter-ferometers (e.g. ALMA in Chile, NOEMA in the French Alps) and by the coarse angular resolution space-borne or ground-basedsurveys (e.g. Planck, ACT, SPT). These instruments have to be installed at the focal plane of the largest single-dish telescopes, whichare placed at high altitude on selected dry observing sites. In this context, we have constructed and deployed a three-thousand-pixeldual-band (150 GHz and 260 GHz, respectively 2 mm and 1.15 mm wavelengths) camera to image an instantaneous circular field-of-view of 6.5 arcminutes in diameter, and configurable to map the linear polarisation at 260 GHz.Aims. First, we are providing a detailed description of this instrument, named NIKA2 (New IRAM KID Arrays 2), in particularfocussing on the cryogenics, optics, focal plane arrays based on Kinetic Inductance Detectors (KID), and the readout electronics.The focal planes and part of the optics are cooled down to the nominal 150 mK operating temperature by means of an ad-hoc dilutionrefrigerator. Secondly, we are presenting the performance measured on the sky during the commissioning runs that took place betweenOctober 2015 and April 2017 at the 30-meter IRAM (Institut of Millimetric Radio Astronomy) telescope at Pico Veleta, near Granada(Spain).Methods. We have targeted a number of astronomical sources. Starting from beam-maps on primary and secondary calibrators wehave then gone to extended sources and faint objects. Both internal (electronic) and on-the-sky calibrations are applied. The generalmethods are described in the present paper.Results. NIKA2 has been successfully deployed and commissioned, performing in-line with expectations. In particular, NIKA2 ex-hibits full width at half maximum (FWHM) angular resolutions of around 11 and 17.5 arc-seconds at respectively 260 and 150 GHz.The noise equivalent flux densities (NEFD) are, at these two respective frequencies, 33±2 and 8±1 mJy ·s1/2. A first successful scienceverification run was achieved in April 2017. The instrument is currently offered to the astronomy community and will remain availablefor at least the following ten years.

Key words. Superconducting detectors – mm-wave – kinetic-inductance – cosmic microwave background – large arrays

1. Introduction

In the past decades progress of astronomical instruments, inparticular the development of large arrays of background-limited detectors, has led to a golden era of millime-

? Corresponding author, e-mail: [email protected]

tre and sub-millimetre continuum astronomy. A number ofinstruments operate hundreds to thousands of very sensi-tive pixels. The majority of these instruments, however,are designed to execute specific scientific programs, mostlikely related to the search of the primordial polarisa-tion modes in the Cosmic Microwave Background (CMB).

1

arX

iv:1

707.

0090

8v4

[as

tro-

ph.I

M]

25

Nov

201

7

Adam et. al.: The NIKA2 large-field-of-view millimetre continuum camera for the 30m IRAM telescope

Few general purpose platforms, like the one described inthis paper, are currently available to the general astronomycommunity. Among them, we cite, for example, Artemis(Reveret et al. 2014) and LABOCA (Siringo et al. 2009) onAPEX (Chile), SCUBA2 (Holland et al. 2013) on JCMT(Hawaii), AzTEC (Chavez-Dagostino et al. 2016) on the LMT(Mexico), MUSTANG2 (Dicker et al. 2014) on GBT (USA),and HAWC+ (Staguhn et al. 2016) on-board SOFIA. Thesecameras are all based on the classical bolometric detection prin-ciple, so far the best approach for this application. In the past tenyears, the Kinetic Inductance Detectors (KID) concurrent tech-nology has demonstrated its competitiveness. For example, thepathfinder NIKA instrument at the IRAM 30-meter telescope,equipped with 356 pixels split over two arrays, has demonstratedstate-of-the art performance in terms of sensitivity, stability, anddynamic range (Monfardini et al. 2010, Monfardini et al. 2011,Catalano et al. 2014, Adam et al. 2014). The most recent ad-vancements in the instrumental domain are described in detailin the LTD16 (Low Temperature Detectors 16) workshop pro-ceedings 1.

Despite the spectacular progression of the technology, sub-millimetre and millimetre studies are often limited by the map-ping speed of intermediate resolution, that is, large single dishes,instruments and their spectral coverage.

Concerning Galactic studies, deep millimetre and sub-millimetre observations at high angular resolution, in intensityand in polarisation, are needed to better understand how starformation proceeds in the interstellar medium (ISM). Solar-typestars form within regions of cold gas in the ISM. These molecu-lar clouds are characterised by an intricate filamentary structureof matter, which hosts the progenitors of stars, that is, pre-stellarand proto-stellar cores (Andre et al. 2010, Konyves et al. 2015,Bracco et al. 2017).

A second key subject is represented by the study of nearbygalaxies which aims at separating physical components by dis-sentangling the emission (e.g. thermal dust, free-free, or syn-chrotron). This allows one to precisely measure the star forma-tion rate in different environments and regions. In this frame-work, the large instantaneous field-of-view of NIKA2 is an as-set.

On the cosmological side, existing CMB experi-ments have proven to be very efficient in detecting clus-ters of galaxies via the Sunyaev-Zel’dovich (SZ) effect(Planck I, 2016, Hasselfield et al, 2013, Reichardt, et al, 2013)and have provided the best cluster cosmological results todate (Planck II, 2016, Planck III, 2016). However, their poorangular resolution limits the cosmological interpretation of thedata and in particular the study of the complex intra clustermedium (ICM) physics, which may bias the observable tocluster-mass scaling relation (Planck III, 2016). This bias mightbe of particular importance for high-redshift, that is, early-stage,galaxy clusters. The dual-band capability of NIKA2 is, in thiscase, crucial.

Similarly, distant universe studies via deep surveys willbenefit from a large instantaneous field-of-view and sensi-tivity to cover sky regions at the confusion limit. This re-sults in detecting dust-obscured optically faint galaxies duringtheir major episodes of star formation in the early universe(Bethermin et al. 2017, Geach et al. 2017).

NIKA2 and the IRAM 30-meter telescope represent todayideal tools to address these scientific questions, and many oth-

1 LTD16, Grenoble, July 20−24th 2015, Journal of Low TemperaturePhysics 184, numbers 1/2 and 3/4, 2016.

ers. The fundamental characteristics of NIKA2 are dual colorand polarisation capabilities, high sensitivity, high angular res-olution and an instantaneous field-of-view of 6.5 arc-minutes.Besides the intrinsic scientific impact, NIKA2 represents the firstdemonstration of competitive performance using large-format(i.e. thousands of pixels) Kinetic Inductance Detector (KID(Day et al. 2003, Doyle et al. 2010) cameras operating at mil-limetre or sub-millimetre wavelengths.

In Section 2 we describe the overall instrument design, in-cluding cryogenics, focal planes, optics and readout electronics.In Section 3 we discuss the observing methods, concentrating inparticular on the photometric calibration procedures. We then, inSections 4 and 5, present the results from the intensity commis-sioning runs at the 30-meter telescope.

2. The NIKA2 Instrument

NIKA2 is a multi-purpose tool able to simultaneously image afield-of-view of 6.5 arcminutes in diameter at 150 and 260 GHz.When run in polarimetric mode, it maps the linear polarisationat 260 GHz. In order not to degrade the native angular resolutionof the 30-meter telescope, and at the same time cover a largefield-of-view, it employs a total of around 2,900 detectors splitover three distinct monolithic arrays of KID. In this Section, wedescribe the main instrument sub-systems and the in-laboratorycharacterisation procedures.

2.1. The cryostat

In order to ensure optimal operation of the detectors and min-imise the in-band parasitic radiation, the focal plane arrays, andthe last portion of the optics, are cooled down to a base tem-perature of around 150 mK by means of a dilution fridge. Thebase temperature must, in fact, be roughly one order of mag-nitude lower than the thin-aluminium superconducting criticaltemperature. The home-made dilution insert is completely inde-pendent and compatible with any cryostat providing a stable 4 Ktemperature input and suitable mechanical and fluid attachmentpoints. We stress that no recycling is needed, the hold time be-ing, in principle, infinite. The dilution refrigerator, and the restof the cryostat, has been entirely designed and realised by CNRSGrenoble. NIKA2 employs two Cryomech PT415 pulses-tubes,each delivering a cooling power of 1.35 W at the reference tem-perature of 4.2 K (second stage) and several tens of Watts at 30-70 K (first stage). The base temperature of these machines is ofthe order of 3 K, sufficient to start the isotopic dilution process.The large cooling power available on the pulse tube’s first stagesallows the integration of a part of the optics baffles at tempera-tures between 4 and 30 K within the cryostat. A cross-section ofthe cryostat is shown in Fig. 1 to illustrate the different cryogenicstages.

Gas heat exchangers are adopted at both pulse-tubes stagesto ensure good thermal contact, avoiding at the same time directmechanical contact between the vibrating parts and the sensitiveinner components, that is, the detectors and cold electronics. Anexternal mechanical regulation of the pulse-tube positions allowsthe optimisation of the cooling power and at the same time theminimisation of the shaking of the coldest components.

The whole instrument is made of thousands of mechanicalpieces, with a total weight of around 1.3 tons when assembled.The weight of the 150 mK stage is of the order of 100 kg, in-cluding several kg of high-density polyethilene (HDPE) low-conductance lenses. Radiation screens are placed at 1 K (still of

2


Fig. 1. (Colour online) Cross-section of the NIKA2 instrumentillustrating the different cryogenic stages. The total weight of thecryostat is close to 1,300 kg. The 150 mK section includes thearrays, the dichroic, the polariser and five high-density polyethi-lene (HDPE) lenses.

the dilution fridge), 4-8 K (pulse-tube second stages), 30 K and70 K (pulse-tube first stages).

Selected inner parts, at each stage of temperature, are coatedwith a high-emissivity mixture of black STYCAST 2850, SiCgrains, and carbon powder. This coating has demonstrated itseffectiveness at millimetre wavelengths in suppressing unwantedreflections (Calvo et al. 2010).

Following the experience accumulated with NIKA, magneticshielding is added on each cryogenic stage, employing high-permittivity materials down to 1 K, and a pure aluminium su-perconducting screen cooled to 150 mK enclosing the detectors.The screening is needed in order to suppress: a) the Earth’s mag-netic field and its variations, in the instrument reference frame,during the telescope slews in azimuth; and b) the magnetic fieldvariations induced by the antenna moving in elevation.

The operation of NIKA2 does not require external cryogenicliquids. Even the needed cold traps are included in the cryo-stat and cooled by the pulse-tubes themselves. The cleaned he-lium isotopic mixture is condensed into liquid when the systemis cold, but it remains in a closed circuit. The cryostat and re-lated sub-systems are fully remotely controlled. The whole cool-down procedure, largely automated, lasts about five days. Fourdays are required for the pre-cooling and thermalisation of thethree coldest stages (see Fig. 1) at around 4 K. During the last24 hours, the dilution procedure is started, allowing the furthercooling down to base temperature. Two additional days before,stable observations are usually foreseen in order to ensure theperfect thermalisation of all the low-thermal-conductance opticselements, such as the lenses, filters and baffle coating. The sys-tem is designed for continuous operations and long observationalruns. So far the base temperature has shown the required stabilityover roughly one month, with no signs of degradation in perfor-mance. The stability of the detector’s temperature is better than0.1 mK RMS over the duration of a typical observational block(scan), that is, roughly 15 min. The common-mode effect on thedetectors of such temperature variations is in this way at leastone order of magnitude lower than the atmospheric fluctuations,even assuming the very best observing conditions. We stressin this context that the maximum gradient between the coldestcryostat parts and the farthest KID array is around 20 mK.

2.2. The focal plane arrays

Each array is fabricated on a single 4” high-resistivity siliconwafer, on which an aluminium film (t = 18 nm) is deposited by e-beam evaporation under ultra-high vacuum conditions. The useof thin superconducting films has a double advantage. First, itincreases the kinetic inductance of the strip, making the detectorsmore responsive, and second, it allows, through its normal statesheet resistance, an almost perfect match of the Lumped ElementKinetic Inductance Detector ( LEKID) meander to the free spaceimpedance of the incoming wave, ensuring a quantum efficiencyexceeding 90% at the peak. The NIKA2 pixels are all based onthe Hilbert fractal geometry that we proposed some years ago(Roesch et al. 2012).

In NIKA, we adopted more classical pixels coupled to a co-planar waveguide (CPW) readout line, with wire bonds acrossthe central line to suppress the spurious slotline mode. Purelymicrofabricated bridges were developed as well. The slotlinemode is associated to a symmetry-breaking between the groundplanes on both sides with the central strip. To optimise the opti-cal coupling to the incoming millimetre radiation, we adopteda back-illumination configuration, in which the light passesthrough the silicon wafer before reaching the pixels. To atten-uate the refraction index mismatch, we micro-machined a gridof perpendicular grooves on the back-side of the wafer, resultingin an effective dielectric constant which is in between vacuumand silicon (Goupy et al. 2016). The total thickness of the siliconwafer, and the depth of the grooves, were chosen to optimise theanti-reflection effect. To maximise the in-band radiation absorp-tion, a superconducting lid was then set at an optimised distancebehind the detector plane, as a λ/4 backshort.

The same approach was originally planned for NIKA2.During the phase of the detector’s development, however, werealised the practical limitations of the CPW coupling approach,in particular considering the thousands of bonds required to en-sure the exclusive propagation of the CPW mode. We then de-cided to study and optimise a different kind of transmissionline, the microstrip (MS). This kind of feedline only supportsone propagating mode, and is thus immune to the risk of spuri-ous modes. Furthermore, the aluminium ground plane is locatedon the opposite side of the wafer with respect to the detectors.This might reduce the still poorly understood residual electro-magnetic cross-coupling between resonators (pixels).

The MS propagation mode shows an electric field oscillatingin the dielectric substrate, between the strip line (feedline) andthe underlying ground plane. This is illustrated in Fig. 2. Themain drawback of the MS coupling lies in the fact that it forces,at least for dual-polarisation imaging applications, front illumi-nation of the detectors. It is thus more adapted for relativelynarrow-band (e.g. ∆ f / f ≤ 30%) applications. This is howeverperfectly compatible with the NIKA2 goals.

In both cases (CPW and MS), the distance between the pix-els and the feedline is chosen in order to satisfy optimal couplingconditions. These are achieved when the coupling quality factor,Qc, is of the same order as the internal quality factor Qi ob-served under typical loading conditions. In the case of NIKA2,we found an optimum at Qc ∼ 10, 000. A metal loop is addedaround each MS-coupled pixel to shield them from the feed-line and achieve the wanted coupling without compromising thecompactness of the pixel packaging (see Fig. 3).

In NIKA2, the 150 GHz channel is equipped with an arrayof 616 pixels, arranged to cover a 78 mm diameter circle. Eachpixel has a size of 2.8 × 2.8 mm2. This is the maximum pixelsize that can be adopted without significantly degrading the the-

3


Fig. 2. Schematic cut through a KID substrate. In grey, the high-resistance silicon wafer, while in black, the aluminium filmsare represented. The green arrows illustrate the direction of theelectric field. Left: the co-planar waveguide (CPW) transmis-sion line without across-the-line bondings, associated to stronglynon-uniform performance of the detector array. Centre: CPWwith across-the-line bondings, configuration adopted in NIKA.Right: the microstrip (MS) configuration adopted in NIKA2, en-suring single-mode propagation and easiest implementation ofvery large arrays.

oretical telescope resolution, as it corresponds roughly to a 1Fλsampling of the focal plane. The array is connected over four dif-ferent readout lines, and shows resonance frequencies between0.9 and 1.4 GHz. The thickness of the silicon substrate is around150 microns, to ensure a maximal optical absorption at 150 GHz.

Fig. 3. Left: a front-illuminated microstrip (MS) pixel for the260 GHz band of NIKA2. The pixel size is 2 × 2 mm2. Right:a back-illuminated coplanar waveguide (CPW) pixel used forthe 150 GHz band in NIKA. The pixel size was in that case2.3 × 2.3 mm2. Both designs are based on Hilbert-shape ab-sorbers/inductors.

In the case of the 260 GHz band detectors, the pixel size is2 × 2 mm2, to ensure a comparable 1Fλ sampling of the focalplane. In order to fill the two 260 GHz arrays, a total of 1,140pixels are needed in each of them. The smaller pixel dimen-sions compared to the 150 GHz band lead to slightly higher reso-nance frequencies that lie between 1.9 and 2.4 GHz. Each of the260 GHz arrays is connected over eight different readout lines.The thickness of the substrate is 260 microns, which maximisesthe optical absorption at 260 GHz. A picture of one of the actual260 GHz arrays mounted in NIKA2 is shown in Fig. 4.

We show in Fig. 5 an illustration of the positioning of thethree arrays in the NIKA2 cryostat.

2.3. The cold optics

In this Section we describe the internal (cooled) optics. More de-tails concerning the telescope interface (room temperature) mir-rors are given in Section 2.6.

NIKA2 is equipped with a reflective cold optics stage heldat a temperature of around 30 K. The two shaped mirrors (M7

Fig. 4. (Colour online) One of the 260 GHz NIKA2 arrays afterpackaging. The number of pixels designed for this array is 1,140,connected via eight feed-lines and 16 SMA (SubMiniature ver-sion A) connectors to the external circuit. The front of the wafercan be seen here.

Fig. 5. (Colour online) Cross-section of the NIKA2 instrumentillustrating the position of the three detector arrays (150 GHz,260 GHz-V and 260 GHz-H). The optical axis and the photondirection of propagation are shown as well.

and M8) are mounted in a specifically designed low-reflectanceoptical box in the cryostat nose. The stray-light suppression isfurther enhanced by a multi-stage baffle at 4 K. The cold aperturestops, at a temperature of 150 mK, are conservatively designedto be conjugated to the inner 27.5 metres of the primary mirrorM1.

The refractive elements of the NIKA2 cold optics aremounted at 1 K and at the base temperature. The HDPE lenses,except for those placed in front of the 260 GHz arrays, are anti-reflecting-coated. The coating is realised by a custom machin-ing of the surfaces. A 30-centimeter-diameter air-gap dichroicsplits the 150 GHz (reflection) from the 260 GHz (transmission)beams. This dichroic, ensuring that it is flatter relative to thestandard hot-pressed ones, was developed in Cardiff specificallyfor NIKA2. A grid polariser ensures then the separation of thetwo linear polarisations on the 260 GHz channel (V and H, seeFig. 5). We refer to Fig. 6 for a schematic cross-section of theinner optics.

The filtering of unwanted (off-band) radiation is provided bya suitable stack of multi-mesh filters placed at all temperature

4


ENTRANCE WINDOW (FROM THE SIDE)

IMAGE STOP& BAFFLE (4K)

APERTURE STOP(150 mK)

DICHROIC(150 mK)

POLARISER (150 mK)

M8 & M7M8 M7

Fig. 6. (Colour online) Cross-section of the NIKA2 instrumentillustrating the cold optics and the main elements and surfacesdescribed in the text. The cold mirrors M7 and M8 are mountedin the cryostat ”nose” on the left side of the Figure.

stages. In particular, three infrared-blocking filters are installedat 300 K, 70 K, and 30 K. Multi-mesh low-pass filters, with de-creasing cutoff frequencies, are mounted at 30 K, 4 K, 1 K andthe base temperature. Band-defining filters, custom-designed tooptimally match the atmospheric windows (see Fig. 10), are in-stalled at the base temperature.

To exploit the NIKA2 polarisation capabilities, a modulatoris added when operating the instrument in polarimetric mode;this consists of a multi-mesh hot-pressed half-wave-plate (HWP)(Pisano et al. 2008) mounted, at room temperature, in front ofthe cryostat window. The modulator uses a stepping motor andis operated at mechanical frequencies of up to 3 Hz, correspond-ing to a maximum of 12 Hz on the effective polarisation modula-tion frequency. A similar setup was successfully used on NIKA(Ritacco et al. 2017). In order to detect all the photons, the mod-ulated polarised signal is then split onto the two 260 GHz arraysby the 45 degree wire-grid polariser described above.

2.4. The readout electronics

One of the key advantages of the KID technology is the simplic-ity of the cold electronics installed in the cryostat. In NIKA2,each block of around 150 detectors is connected to single coax-ial line providing the excitation at one end, and the readout atthe other. The excitation lines, composed of stainless-steel ca-bles, are running from 300 K down to sub-Kelvin temperature.They are properly thermalised at each cryostat stage, and a fixedattenuation of 20 dB is applied at 4 K in order to suppress theroom temperature thermal noise. Each excitation line ends witha SMA (SubMiniature version A) connector (EXCitation in-put) and an ad-hoc launcher connected, through superconducting(aluminium) micro-bonds, to the silicon wafer holding the detec-tors. The approximate excitation power per resonator is typicallyof the order of 10 pW.

On the readout side, the same types of micro-bonds are usedto transfer the signal out of the focal plane and to make it avail-able on a second SMA connector (MEASurement output). Thena superconducting (Nb) coaxial cable is used to connect the mea-surement output directly to the input of a cryogenic low-noiseamplifier (LNA). The amplified signal provided by the LNA istransferred through the remaining cryostat stages (up to 300 K)

PO

WE

R S

UP

PLY

MC

H+

CC

SB

PPS

CLK

ETH

482.6 mm (19'')

311

mm

(7U

)

260.

35 m

m

(10.

5'')

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

NIK

EL

_AM

C

LO

EXC

MEAS

FM

Fig. 7. Overview of one array readout electronics crate. It isequipped with 4 (or 8) readout boards lodged in advancedmezzanine card slots (NIKEL AMC), one central and clockingand synchronisation board (CCSB) mounted on the MicroTCACarrier Hub (MCH) and one 600 W power supply. The crate al-located to the 150 GHz channel uses four NIKEL AMC boardswhile the others use eight NIKEL AMC boards. The actualpower consumption is around 500 W per crate for eight boardsand 300 W for four boards. The weight of each crate is around12 kg.

via stainless-steel coaxial cables. The LNAs, which operate atfrequencies up to 3 GHz, show noise temperatures between 2 Kand 5 K and are held at a physical temperature of about 8 K. Thismeans that the input amplifier noise is equivalent to the thermalJohnson noise of a 50-Ohm load placed between 2 K and 5 K .The cryogenic amplifiers used in NIKA2 were developed, fabri-cated, and tested at the Yebes observatory and TTI Norte com-pany, both located in Spain. The specifications of the amplifiershave been elaborated by the NIKA2 group. In total, NIKA2 iscomposed of about 2,900 pixels and is equipped with twentyfeed-lines. Thus, it employs twenty cryogenics amplifiers (fourfor the 150 GHz array and eight for each of the 260 GHz arrays).The polarisation of the LNA stages is provided by a custom elec-tronics box remotely controlled and allowing the optimisationsof the biases according to the slightly different characteristics ofthe front-end High Electron Mobility Transistors (HEMT).

The warm electronics required to digitise and process the2,900 pixels’ signals were specifically designed for that pur-pose; it is composed of twenty readout cards (one per feed-line) named New Iram Kid ELectronic in Advanced MezzanineAard format (NIKEL AMC). As shown in Fig. 7, the cardsare distributed in three micro-Telecommunication ComputingArchitecture (MTCA) crates. A central module, composed of acommercially available Mezzanine Control Hub (MCH) and ofcustom-made mezzanine boards, is used to distribute a 10 MHzrubidium reference clock (CLK) and a pulse per second (PPS)signal provided by a global positioning system (GPS) receiverand to control the crate. The synchronisation between the boardsand the different crates is ensured by these common referencesignals. The electronics is fully described in previous papers(Bourrion et al. 2012, Bourrion et al. 2016).

In summary, the NIKEL AMC is composed of two parts:the radio-frequency (RF) part and the digitisation and processingsection. The integrated RF part ensures the transition from andto the baseband part. It uses the local oscillator (LO) input toperform up and down conversions. To instrument the 150 GHzarray (resonances from 0.9 GHz to 1.4 GHz) and the 260 GHzarrays (resonances from 1.9 GHz to 2.4 GHz), the used LO input

5


frequencies are 0.9 GHz and 1.9 GHz, respectively. The digiti-sation and signal processing, which is done at baseband, relieson channelised digital down conversion (DDC) and their asso-ciated digital sine and cosine signal generators and processors.The processing heavily relies on field programmable gate arrays(FPGA) while the interfacing to and from the analog domain isachieved by 1 GSPS analog to digital and digital to analog con-verters (ADC and DAC, respectively). The electronics covers abandwidth of 500 MHz and handles up to 400 KIDs in this band-width. In NIKA2 about 150 KIDs per board are instrumented,leaving room for placing a number of dark and/or off-resonanceexcitation tones, and allowing for future developments of theinstrument. It must be noted that for implementation reasons(Bourrion et al. 2012, Bourrion et al. 2016) the excitation signal,nominally covering 500 MHz, is constructed by five DACs, eachspanning 100 MHz.

2.5. Laboratory tests

NIKA2 has been pre-characterised in the laboratory under re-alistic conditions. In order to deal with the absence of thetelescope optics, we have added a corrective lens at the cryo-stat input window. This lens generates an image of the focalplanes onto our ”sky simulator”, as described in previous pa-pers (Catalano et al. 2014, Monfardini et al. 2011). A sub-beam-sized, that is, point-like, warm source, moved in front of the skysimulator by means of an x-y stage, allows beams and array ge-ometry (e.g. pixel-per-pixel pointing) characterisation. The sen-sitivity is calculated by executing calibrated temperature sweepsof the sky simulator, and measuring the signal-to-noise ratio.A photometric model has been developed based on ray-tracingsimulations. The overall transmission of the instrument, mainlydetermined by the lenses and the filters, lies around 35%. Thefilters have been individually characterised in Cardiff, while thelens transmission has been measured at IRAM, Grenoble. Ontop of that, the quantum efficiency of the detectors, integratedin the band of interest and calculated from ab-initio electro-magnetic simulations, is between 60% (260 GHz arrays) and80% (150 GHz array). The three-dimensional (3D) electromag-netic simulations, for the Hilbert design adopted in NIKA2,are confirmed by millimetre-wave vector analyser measurements(Roesch et al. 2012).

The frequency sweep of the four lines connected to the150 GHz array is shown in Fig. 8. The number of identified res-onances over the twenty feedlines exceeds 90% when comparedto the number of pixels implemented by design.

The measurable quantity, proportional to the incomingpower per pixel, is the shift in frequency of each resonance(pixel) (Swenson et al. 2010). This is the reason why our noisespectral densities are expressed in Hz/Hz0.5, and the Rayleigh-Jeans responsivities are given in kHz/K. By sweeping the tem-perature of the sky simulator, we have estimated average respon-sivities around 1 and 2 kHz/K at 150 and 260 GHz, respectively.The array-averaged frequency noise levels for the two bands areabout 1 and 3 Hz/Hz0.5, resulting in noise equivalent temperature(NET) of the order of 1 and 1.5 mK/Hz0.5 per pixel at 150 and260 GHz, respectively. These figures are calculated at a repre-sentative sampling frequency of 5 Hz. An example measurementbased on the sky simulator, and determination of the responsiv-ity, is reported in Fig. 9.

Noise spectra recorded in the laboratory have been fully con-firmed with NIKA2 when operating at the telescope. Please re-fer to Section 4.5 for a more detailed discussion concerning thenoise properties.

Fig. 8. (Colour online) Resonance sweep for the four feedlines ofthe 150 GHz array. The lines 1,2,3,4 are shown in blue, red, cyanand green, respectively. The y-axis represents the transmissionof the feedline (parameter S21) and is expressed in dB. Each dipcorresponds to a resonance/pixel. At least 94% of the 616 pixelsare identified with a resonance and are thus sensitive to incomingradiation.

Fig. 9. (Colour online) Responsivity estimation using the skysimulator. Top panel: frequency sweep of a portion of one par-ticular feedline operating at 260 GHz. Bottom panel: zoom onthree typical resonances. In both panels we plot the S21 trans-mission parameter (dB) against the frequency. Blue lines: coldsky simulator (TS S ∼ 80 K). Red lines: 300 K background. Themeasured average responsivity, that is, the shift in frequency perunit temperature background variation, is around 2 kHz/K for the260 GHz arrays and 1 kHz/K in the case of the 150 GHz array.

The spectral characterisation of the arrays and the overall op-tical chain of NIKA2 (Fig. 10) has been achieved in the Grenoblelaboratory using a Martin-Puplett Interferometer (MpI) built in-house (Durand 2008) and specifically dedicated to characteri-sation. The band transmission in Fig. 10 was measured usinga Rayleigh-Jeans spectrum source. The two arrays operating at260 GHz, mapping different polarisations, exhibit a slightly dif-ferent spectral behaviour, probably due to a tiny difference in thesilicon wafer and/or Aluminium film thicknesses. The observedshift of the peak frequency, 265 GHz for the V (A1) array versus258 GHz for the H (A3), can be explained by an approximately5-micron change in the substrate thickness. The so-called 1 mmatmospheric window is not completely filled. This was designedfor the first generation of detectors in order to ensure robustness

6


against average atmospheric conditions and to optimise the over-all observing efficiency. A possible future upgrade of NIKA2,oriented towards even better sensitivity in very good atmosphericconditions, would be straightforward.

100 150 200 250 300 350Frequency [GHz]

0.0

0.2

0.4

0.6

0.8

1.0

Norm

aliz

ed s

pect

ral Tra

nsm

issi

on

Fig. 10. (Colour online) NIKA2 spectral characterisation for thetwo 260 GHz arrays, H (A1, blue) and V (A3, green) measuredin the NIKA2 cryostat, and for the 150 GHz array (A2, red) mea-sured in a test cryostat equipped with exact copies of the NIKA2band-defining filters. The band transmissions are not correctedfor Rayleigh-Jeans spectrum of the input source. We also showfor comparison the atmospheric transmission (Pardo et al. 2002)assuming 2 mm of precipitable water vapour (PWV), that is, verygood conditions, and 6 mm PWV, that is, average conditions.

The sky simulator enabled also a rough but crucial estima-tion of the parasitic radiation. By comparing measurements ob-tained at several sky simulator distances with respect to the cryo-stat window, we determined an equivalent 15 K additional focalplane background due to the ambient temperature stray radia-tion. This is lower than the very best equivalent sky temperatureat Pico Veleta (≈ 20 K), and confirms that NIKA2 is not signif-icantly affected by this effect. In comparison, in NIKA we hadestimated around 35 K additional background, slightly limitingthe performance. In summary, the overall performance of the in-strument, measured preliminarily in the laboratory, is in line withthe NIKA2 specifications, paving the way for the installation atthe telescope described briefly in the following Section.

2.6. The integration at the telescope

NIKA2 was transported from the Grenoble integration hall to theobservatory at the end of September, 2015. Successful installa-tion of the instrument took place in early October, 2015, at theIRAM 30-meter telescope on Pico Veleta (Sierra Nevada, Spain).To prepare this installation, the optics of the receiver cabin (M3,M4, M5 and M6) had been modified in order to increase the tele-scope field-of-view up to the 6.5 arcminutes covered by NIKA2.M3 is the Nasmyth mirror attached to the telescope elevation

axis. M4 is a flat mirror that can be turned manually in orderto feed the beam either to NIKA2 or to heterodyne spectro-scopic instruments (Carter et al. 2012, Schuster et al. 2004). TheM5 and M6 curved mirrors are dedicated to the NIKA2 cam-era. The configuration of the optics in the cabin, for an elevationδ = 0 degrees, is drawn in Fig. 11. Not shown nor discussed, M1and M2 are the telescope primary mirror and its sub-reflector,respectively.

Fig. 11. (Colour online) Left: Isometric view of the cabin opticsscheme, illustrating the mirrors M3, M4, M5 and M6. The idealcase in which the elevation angle is zero degrees is shown. Right:Top view of the cabin optics feeding NIKA2.

The whole installation, including the cabling of the instru-ment, was completed in about three days. The pulse-tube pipes,which are 60 metres long, run through a derotator stage in orderto connect the heads in the receiver cabin (rotating in azimuth)and the compressors located in the telescope basement (fixed).A single 1 Giga-bit ethernet cable ensures the communicationto and from the NIKA2 instrument. The forty radio-frequencyconnections (twenty excitation lines, twenty readouts) betweenthe NIKEL AMC electronics and the cryostat, located on op-posite sides of the receivers cabin, are realised using 10-meter-long coaxial cables exhibiting around 2 dB signal loss at 2 GHz.This is acceptable, considering that the signal is pre-amplifiedby about 30 dB by the LNAs.

The optical alignment between the instrument and the tele-scope optics has been achieved using two red lasers. The firstwas set shooting perpendicularly from the centre of the NIKA2input window, through the telescope optics and reaching the ver-tex and M2. The second laser was mounted on the telescope el-evation axis at the M3 position, reaching then, through the M4,M5 and M6 mirrors, the NIKA2 window. In both cases, we haveadjusted the cryostat position and tilt. NIKA2 is equipped withan automatic system of pneumatic actuators and position detec-tors able to adjust the cryostat height and tilt and to keep it stabledown to a few tens of microns precision.

The first cryostat cooldown started immediately after the in-stallation, and was achieved after the nominal four days ded-icated to pre-cooling, followed by less than 24 hours duringwhich the helium isotopes mixture is condensed in the so-called”mixing chamber”. The first-light tests demonstrated that all thedetectors were functional and exhibited responsivity and noise inline with the laboratory measurements presented in the previousSection. The preliminary results of the initial technical runs arepresented in ?.

3. Measurement principle

At the telescope, the NIKA2 acquisitions on a given source aresplit into single observational blocks referred to as ”scans”. In

7


Fig. 12. (Colour online) A picture of the NIKA2 cryostat in-stalled in the 30-meter telescope receivers cabin. Picture takenin October, 2015.

each scan, the source is moved across the field-of-view, typicallyby constant-elevation pointing sweeps defined as ”sub-scans”. Aparticular scan, in which the elevation of the antenna is movedby large amounts, is named ”skydip”. The skydip allows us tomeasure the effective atmosphere temperature and calibrate thesky opacity corrections, as explained in Section 3.3.

The photometry can be reconstructed down to the requiredprecision thanks to three distinct procedures that are applied dur-ing sky observations. First, we implement a real-time electricalcalibration acting directly on the KID that is specific to NIKAand NIKA2 (Sect. 3.1). Second, KID measurements are dynam-ically adapted to the sky background (Sect. 3.2). Third, the at-mosphere opacity correction is calculated in real time thanks tothe large dynamic range and linearity of the detectors (Sect. 3.3).

3.1. Internal electrical detector calibration

When radiation is absorbed in a KID, it breaks part of the su-perconducting carriers (Cooper pairs) and creates a non-thermalexcess of unbound electrons (quasi-particles). This changes theimpedance of the film and shifts the resonance frequency fr ofthe KID to lower values. The standard way to read a pixel (res-onator) is to excite it with a tone at a frequency ft and monitorhow the in-phase (I) and in-Quadrature (Q) components of thetransmitted signal are modified by the changes in its resonancefrequency fr. For NIKA2 we adopted the strategy already devel-oped and tested in NIKA. Instead of using an excitation at a fixedfrequency ft, we rapidly ( fmod ≈ 500 Hz) modulate between twodifferent readout tones, f +

t and f −t , ideally placed just above andjust below fr. The tones are separated by d f = f +

t − f −t , muchsmaller than the resonance width. This modulation technique al-lows us to measure, for every data sample, both the values ofI and Q and the variation dI, dQ that is induced by the cho-sen frequency shift d f . When the optical power on the detectorschanges by an amount ∆Popt, a variation ∆I, ∆Q is observed be-tween successive data samples, which are acquired at a rate offsampling = 24 ÷ 48 Hz � fmod. The dI, dQ values can then beused as a calibration factor to associate to the observed ∆I, ∆Q, the corresponding change in the resonance frequency ∆ fr, andthus measure ∆Popt. A full description of the modulated readouttechnique is provided in Calvo et al. 2013.

The advantage of this solution is that the dI, dQ values areevaluated for every data sample. If the load on the detectorschanges (e.g. due to variations in the atmosphere opacity), the

exact shape of the resonance feature of each pixel will change.However, since the calibration factor dI, dQ is updated in realtime, this effect is taken automatically into account. The photo-metric accuracy of the instrument, through the KID linearity interms of fr, is thus strongly improved.

Furthermore, knowing both the I, Q and the dI, dQ valueswe can also estimate the difference between fr and ft. In theideal situation, these two frequencies should coincide. In reality,changes in the background load can make the resonances driftby a large amount. In such a case, the modulated readout allowsus to rapidly readjust ft to the instantaneous value of fr. Thisensures an optimal frequency bias and prevents any degradationin the sensitivity of the detectors. This returning is scheduled, asdiscussed in more detail in Sect. 3.2, between different scans orsub-scans and does not affect the data taken during the properintegration.

3.2. Sky background matching

During ground-based observations, the radiation load per pixelis determined by the atmospheric transmission and pointing el-evation. The load itself is variable in time due to the atmo-sphere opacity fluctuations around its mean value. The KIDtone-frequency load-matching procedure, which we call ”tun-ing”, is performed in a specifically dedicated sub-scan at the be-ginning of each scan and in the lapse of time between two subse-quent scans. The tuning procedure is usually performed as a two-step process. First, a common shift is applied to all KIDs in orderto match the instantaneous average sky background. Second, theKIDs are individually adjusted by fine-tuning their position. Thetwo steps, depending on the weather conditions, can be executedseparately. The versatility of the tuning procedure allows us tokeep track of the KID resonance positions even under variableobserving conditions, or when the elevation is changed strongly,for example during skydips. The complete tuning, including averification of the correct frequencies adjustment, is completedin less than 2 seconds.

The tuning procedure requires real-time synchronisation ofthe NIKA2 camera with the telescope control system. This isachieved by directly receiving and interpreting the telescope sta-tus messages. These messages are broadcast by the telescopeserver over the NIKA2 private network at a rate of 8 Hz. The in-terpreted messages (e.g. begin and end of scans and sub-scans)are recorded in the NIKA2 raw data files. In addition, off-lineaccurate (< 0.1 msec) synchronisation of the telescope attitudefile and the NIKA2 raw data is obtained by monitoring the PPS(Pulse Per Second) signal. This signal, as mentioned in Sect. 2.4,is generated by the telescope control system and shared by all theinstruments.

3.3. Atmospheric attenuation correction

The sky maps have to be corrected for the atmospheric absorp-tion. The corrected brightness S corrected is:

S corrected = S ground · ex·τscan , (1)

where τscan is the zenith opacity, for each band, of the atmo-sphere during the observation, and x represents the airmass2 atthe considered elevation.

2 By assuming a homogeneous plane-parallel atmosphere, the rela-tion between the airmass and the elevation of the telescope is taken asx = csc(δ) = 1/sin(δ), where δ is the average elevation.

8


Fig. 13. (Colour online) Atmospheric opacity as measured fromthe IRAM 225 GHz tau-meter (cyan), and from the NIKA2data at 150 (red) and 260 GHz (blue) during the February 2017NIKA2 commissioning campaign. We stress the fact that theIRAM 225 GHz tau-meter data is not used for the atmosphericcorrection and is plotted here just for comparison.

NIKA2 has the ability to compute the opacity directly alongthe line of sight, integrated in NIKA2’s exact bandpasses andindependently of the IRAM tau-meter operating at 225 GHz. Theprocedure was successfully tested with NIKA and is describedin detail in Catalano et al. 2014. Indeed, the resonance frequencyof each detector is related to the airmass and the opacity as in thefollowing:

f ir = ci

0 − ci1 · Tatm[1 − e−x·τ], i = 1...N, (2)

where f ir is the measurable absolute value of the ith detector res-

onance frequency (Sect. 3.1), ci0 is a constant equal to the reso-

nance frequency at zero opacity, ci1 is the calibration conversion

factor in kHz/K, with Tatm we refer to the equivalent tempera-ture of the atmosphere (taken as a constant at 270 K), τ is thesky zenith opacity, as a function of time, and N is the number ofuseful detectors in the considered array.

The coefficients ci0 and ci

1 are expected to be constant in timewithin at least one cooldown cycle; once they are known, Eq. (2)can be inverted to determine τ at the corresponding time. The de-termination of the constants ci

0 and ci1 is achieved via a specific

scan, nicknamed skydip. For such a scan, the telescope performseleven elevation steps in the range δ = 19 ÷ 65 deg, regularlyspaced in airmass. For each step, we acquire about twenty sec-onds of time traces to reduce the error in the determination of f i

r .Several skydips under different weather conditions (hence dif-ferent τ) are solved simultaneously in order to break the naturaldegeneracy between the opacity and the responsivity.

We observe that the skydip-fitted τ values are, as expected,common between different detectors of the same array. By com-paring the results of different skydips, we have verified exper-imentally that the coefficients ci

0, ci1 are stable, within the fit

errors, on very long time scales within a cooldown cycle. Thecoefficients can thus be applied to the whole observing cam-paign in order to recover the opacity of each scan. In Fig. 13 wepresent the evolution of the NIKA2 in-band opacities for severalscans of the commissioning run held in February 2017. Theseare compared to the IRAM tau-meter readings. We observe aglobal trend agreement between the IRAM-tau-meter-suggested

opacity (225 GHz) and the NIKA2 values. These latter valuesshow, however, a smaller dispersion. We find an average ratiobetween the 150 GHz and the 260 GHz NIKA2-derived opaci-ties of about 0.6, which is only broadly consistent with modelexpectations. We notice however that the 150 GHz-to-260 GHzopacity ratio varies significantly for opacities (at 150 GHz) be-low 0.2. This effect is likely to be linked to an O2 atmosphericline which becomes saturated. This point is, however, still underinvestigation.

4. Observations and performance

The first NIKA2 astronomical light was achieved in October2015. A first technical run immediately followed. A number ofcommissioning runs were then carried out between November,2015, and April, 2017. The commissioning observations werecarried out periodically and did not interfere with the hetero-dyne routine observations scheduled at the telescope. SinceSeptember, 2016, the instrumental configuration has been fullystable after replacing the 150 GHz array, the dichroic, andthe smooth lenses with anti-reflection ones. In this paragraphwe summarise the first results obtained for the characterisa-tion of the instrument performance. More in-depth details con-cerning the data analysis pipeline and the commissioning re-sults will be given in forthcoming papers (Ponthieu et al. 2017and Perotto et al. 2017). The commissioning of the polarisationchannel is on-going, and results will also be the object of a futurepaper.

We stress that the experience from the use of NIKA2 by ex-ternal astronomers might lead, in the best case, to further op-timisation of the instrument performance. The experience thatwill be accumulated in the future might eventually allow us toevaluate subtle problems that have not been solved during thecommissioning.

4.1. Data processing at the telescope

Astrophysical observations are carried out in scans, whichtypically last for a few minutes up to 20 minutes at most.The NIKA2 data in intensity are sampled at 23.8418 Hz. Inpolarimetric mode, the sampling runs at twice this frequency,that is, 47.6836 Hz. For each sample and for each KID, theIn-phase (I), the Quadrature (Q) and their derivatives (dI, dQ) ofthe transfer function of the feed-line and the pixel are recorded.Some extra information like scan number, sub-scan number, andtelescope pointing information are also included.

Data processing is needed during telescope observations toensure the scientific quality of the acquired data. We have devel-oped two sets of tools for real-time and quick-look analysis. Thereal-time tools run on a dedicated multi-processor acquisitioncomputer and monitor the KID time ordered data and the overallbehaviour of the instrument, and decode telescope messages asdiscussed in Sect. 3.2.

The quick-look analysis software is run at the end of eachscan by the observer to obtain early feedback. The executiontime is typically of the order of one minute. It includes both mapmaking capacities and analysis of the produced maps, in termsof photometry, sensitivity and calibration. Indeed, it is used tomonitor the pointing of the telescope and its focus (about onceevery two hours in normal observing conditions) and to give therequired instructions to correct for their drifts. This quick-look

9


software has been used extensively during the commissioning.It shares a number of common tools with the off-line processingpipeline that are used to construct the final and optimal sky maps(Ponthieu et al. 2017).

4.2. Field-of-view reconstruction

The reconstruction of the position of the detectors in the field-of-view (FoV) is mainly based on observations of planets, andin particular Uranus, Neptune, and Mars. We generally performdeep-integration azimuth raster-scan observations at constant el-evation. A total of 99 subscans are taken by changing elevation insteps of 4.8′′. The overall footprint of these scans, which we callbeam-maps, is 780′′ ×470′′. We produce a map of the source foreach KID with a projected angular resolution of 4′′. These mapsare used to derive the KID position on the FoV, the propertiesof the beam pattern (FWHM and ellipticity) per KID, and thedetector inter-calibration.

Figure 14 shows the position of the detectors in the NIKA2FoV for the two 260 GHz arrays (A1 and A3), and for the150 GHz array (A2). For each detector the ellipse symbol sizeand ellipticity are proportional to the main beam FWHM and el-lipticity, as derived from a fit to a 2D elliptical Gaussian. To iso-late the main beam contribution to the total beam, the sidelobesare masked out using annulus masks centered on the peak signal,of 50′′ external radius and of internal radius of 9′′ at 260 GHzand 14′′ at 150 GHz. Elliptical 2D Gaussian fits on the maskedmaps of each detector (individually) provide two orthogonal-direction FWHMs, which are geometrically combined to obtainthe main beam FWHM. Figure 15 shows the distribution of themain beam FWHMs of the arrays A1, A3, and A2 using a beam-map scan of Neptune acquired during the April 2017 commis-sioning campaign and for average weather conditions. We alsoshow in red the best Gaussian fit to histogram data. We find anaverage main beam FWHM of 10.9′′ at 260 GHz and 17.5′′ at150 GHz in agreement with the main beam estimates from thedeep beam map presented in Fig. 16. The observed dispersion ofabout 0.6′′ is expected from the optics design and its associatedfield distortions across the FoV of 6.5 arc-minutes. In particular,these distortions lead to an optimal focus, determined by the po-sition of the M2 mirror, that is shifted by 0.2 mm between thecentre and the outer edge of the FoV.

4.3. Antenna diagram

To probe the extended beam patterning, we use the same ob-servations as above, and produce a map with all usable detec-tors. We show in Fig. 16 the beam pattern as obtained fromthe optical instrument and telescope response to Uranus for ar-rays A1, A3, the combination of A1 and A3 (260 GHz), andA2 (150 GHz). The telescope beam is characterised by its mainbeam, side lobes, and error beams. The main beam is well de-scribed by a 2D Gaussian, while the error beams are more com-plex and have been fitted to the superposition of three Gaussiansof increasing FWHM (65′′, 250′′ and 860′′ at 210 GHz) inGreve et al, 1998, Kramer et al, 2013 using observations of thelunar edge with single pixel heterodyne receivers to constructbeam profiles.

The maps are consistent with a 2D gaussian main beam ofFWHM 11.3′′ ± 0.2′′, 11.2′′ ± 0.2′′, and 17.7′′ ± 0.1′′, for A1,A3 and A2 arrays, respectively. These results are consistent withthe mean FWHM of the main beam of individual KIDs shown inFig. 15 and are presented in Table 1. The colour lines in Fig. 16

show the relatively complex sidelobes and error beams. In the260 GHz maps (A1, A3, and combined A1 & A3) we clearlyobserve a diffraction ring at a radius of about 100′′ and at -30dB. The diffraction ring and its spokes are presumably causedby radial and azimuthal panel buckling (Greve et al, 1998). Theperpendicular green lines shown in the A2 (150 GHz) map cor-respond to the diffraction pattern caused by the quadrapod struc-ture supporting M2. In the same map the yellow arrows pointto four symmetrical spokes of the error beams. The pink el-lipses show spikes in the A2 map. We observe, in the A3 mapin Fig. 16, some spikes of unknown origin.

Comparing the 2D Gaussian main beam fit to the full beampattern measurement up to a radius of 250′′, we compute thebeam efficiencies defined as the ratio of power between the mainbeam and this full beam. We find beam efficiencies of ∼ 55 %and ∼ 75 % for the 260 and 150 GHz channels, respectively.Heterodyne observations of the lunar edge and of the forwardbeam efficiency derived from skydips show that a significantfraction of the full beam is received from beyond a radius of250′′. This fraction is not considered here.

4.4. On-sky calibration

The planets Uranus and Neptune were used as the primary cal-ibrators. Their reference flux densities were obtained from themodel in Moreno, 2010, Bendo et al. (2013) and updated at themid-date of each session of observations. We use the planetgeocentric distance and viewing angle to account for planetaryoblateness as provided by the JPL’s HORIZONS Ephemeris3.By convention, flux densities are given at reference frequencies150 and 260 GHz for the two channels, respectively. We noticethat the reference frequencies are close to the peak frequenciesfor each channel, which were discussed in Sect. 2.5 .

Various observations of Uranus and Neptune with integra-tion times of ∼ 20 minutes were carried out during each com-missioning session resulting in high SNR maps (e.g. Fig. 16).These observations of strong sources with relatively long inte-gration times were planned to minimise statistical noise in orderto determine the level of systematics that characterises the sta-bility of the scale. Their total flux densities were measured fromthe maps by aperture photometry within a radius of 150′′ wherecumulative flux density levelled off smoothly. For this photome-try, we used the solid angle of the beam of the telescope that wecould determine for each observation with these strong sourcesin using their radial brightness profile computed at the maxi-mum extent rmax = 250′′. The beam solid angle was found tobe slightly variable, as expected since no telescope gain depen-dence on elevation was yet implemented for the processing of theNIKA2 observations, but also because atmospheric conditionsmight have had an impact. The obtained fluxes were correctedfor atmospheric absorption using the atmospheric line-of-sightopacity for the two NIKA2 channels, which was computed asdescribed in 3.3.

The ratios between the fluxes for each individual planet ob-servation and the reference planet flux density for the threeNIKA2 arrays are shown in Fig. 17 for the February and April,2017, commissioning campaigns. The mean ratios for the threearrays are close to unity as expected since the planets were usedto set the calibration factors in the off-line processing. Overall,the flux density scale is stable at better than 7% for all observa-tions acquired during the two one-week runs (separated by twomonths) and despite the fact that the instrument was warmed

3 https://ssd.jpl.nasa.gov/horizons.cgi

10


up in between the two sessions. It is noticeable that the scat-ter around unity in Fig. 17 is about twice smaller in the first(February) session, which was conducted in significantly betterweather conditions.

More precisely, we quantify the stabilities for the Februarycommissioning week via the relative flux rms and find values of3.6%, 2.5% and 2.9% for arrays A1, A2 and A3, respectively,with atmospheric opacity at 260 GHz between 0.05 and 0.3.Correspondingly, for the April campaign, we find 5.3%, 6.7%and 8.6% with atmospheric opacity at 260 GHz between 0.25and 0.6. It is thought that, at the moment, limitations in stabilityare caused by residual atmospheric fluctuations and uncertaintyin opacity corrections.

Nonetheless, the flux density scale of NIKA2 is found to behighly stable and comparable to the level achieved by other mod-ern instruments, such as SCUBA2 (Dempsey et al. (2013)). Theother limitation of the scale is absolute calibration that dependson the accuracy of the Moreno, 2010, Bendo et al. (2013) modelwhich is estimated to be 5% in the millimetre wavelength range.Hence, in combining both limitations, the total uncertainty ofcalibration with NIKA2 is 10% in mediocre atmospheric condi-tions and better than 6% in fair conditions (τ260GHz < 0.3).

4.5. Noise and sensitivity

We have investigated the noise properties and sensitivity ofNIKA2 in various atmospheric conditions and for various typesof sources including both faint and bright ones. For each obser-vation scan the raw data were corrected for atmospheric fluctu-ations, which are seen as a common mode by all or most of thedetectors. These corrected data are then projected into maps.

Figure 18 shows an example, in typical weather conditionsτ260 GHz < 0.3, of the power spectrum of the NIKA2 time or-dered data before (black) and after (blue) subtraction of theatmospheric fluctuations. We observe that even in the case ofgood weather conditions the signal is dominated by atmosphericfluctuations, in particular at small frequencies, giving a 1/f-likespectrum. After atmospheric subtraction and proper filtering, thespectrum is flatter. A detailed study of the residual correlatednoise, in both the time-ordered data and maps, will be given inthe companion papers Perotto et al. 2017, Ponthieu et al. 2017.

The NEFD is routinely estimated on those maps for eachindividual scan from the measured flux uncertainties across themap. Using the NEFD we obtain the mapping speed as ms =ε ·FOV/NEFD2 where FOV is the field-of-view solid angle andε is the fraction of used detectors. We stress that the NEFD val-ues assume the spectral energy distribution (SED) of the primarycalibrators.

From our analysis we find that the measured NEFDs per scanare consistent with being background dominated both for the 150and 260 GHz NIKA2 channels. We observe some residual cor-related noise in the per scan maps mainly due to residual atmo-spheric contamination. However, when averaging across scansthe noise evolves consistently with the square root of the time ofobservation over more than 3 hours of observations. We achievedaverage sensitivities, extrapolated at τ = 0 from a fit (to be easilycomparable to other instruments) of 8 and 33 mJy·s1/2 at 150 and260 GHz, respectively. The fit versus tau was performed basedon a dataset including observations made under airmasses in therange from 0.05 to 0.85. These NEFDs correspond to mappingspeeds of around 1350 and 73 arcmin2/hr/mJy2. These values re-fer to the average over many scans taken under different observ-ing conditions. They are conservative with respect to the valuesobtained in the best scans under good weather conditions, that

is, 2 mm PWV and elevation δ = 60◦. We have thus gainedan order of magnitude mapping speed over the previous gener-ation of instruments at the 30-meter telescope (i.e. NIKA andGISMO operating at 150 GHz, NIKA and MAMBO2 operatingat 260 GHz).

4.6. Summary of performances

We present in Table 1 a summary of the main characteristics andperformance of the NIKA2 instrument for the February com-missioning campaign. From this table we conclude that NIKA2behaves better than the initial goals at 150 GHz, and is compat-ible with the specifications at 260 GHz (Calvo et al. 2016). TheNEFD sensitivity at 260 GHz is limited by sky noise decorre-lation techniques (Perotto et al. 2017, Ponthieu et al. 2017) anda still unidentified optical problem reducing considerably the il-lumination on the array A1 (260 GHz-H). This issue is underinvestigation and will be addressed in a forthcoming refurbish-ment.

5. Illustration of NIKA2 mapping capabilities

During commissioning and the science verification phase weobserved several compact and extended sources in order tocheck the NIKA2 mapping capabilities. Here we concentrateon two sources to illustrate the main advantages of NIKA2with respect to previous experiments. A more detailed descrip-tion of the sources observed will be given in companion papers(Perotto et al. 2017, Ponthieu et al. 2017).

In Fig. 19, we present 260 (top) and 150 (bottom) GHzNIKA2 maps of the star system MWC349, which is awell known secondary calibrator at millimetre wavelengths.MWC349 was systematically observed at different elevationsand in different weather conditions during the February andApril, 2017, commissioning campaigns to monitor the stabilityof the calibration. In the Figure, we present the averaged mapobtained using all scans taken during the February campaignfor a total integration time of 3.44 hours. We clearly observeMWC349 in the centre of the two maps with high significance.The measured MWC349 fluxes are given in Table 2. We observetwo other sources at the edges of the maps, which we refer to asSO1 and SO2. Details on the position and fluxes of the sourcesare given in Table 2. The SO2 source corresponds most prob-ably to the radio-millimetre source BGPS G079.721+00.427,which was detected by Bolocam in their Galactic plane sur-vey (Rosolowsky et al. 2010). In the Bolocam observations, theSO1 and SO2 sources cannot be distinguished from each other.Furthermore, the SO1 source might correspond to KMH2014J203310.31+404118.72 (Kryukova et al. 2014), a young stellarobject, but the identification is not sufficiently secure. From theseobservations we conclude that the large FOV and high sensitiv-ity of NIKA2 translate into a large mapping speed that allows usto cover a large sky area with the possibility of observing, and/ordiscovering, various sources simultaneously.

To test the ability of NIKA2 to detect faint sources, we per-formed a concerted series of observations of Pluto and Charon.During these observations the atmospheric opacity was stable:about 0.25 at 260 GHz. The maps are shown in Fig. 20. Forthis paper, we concentrate only on the central region of themaps where we observe a significant detection of the Pluton andCharon planetary system. To illustrate this, on the image, wehave also superimposed signal-to-noise ratio contours at valuesof 5, 7, and 10. The fluxes and uncertainties of the Pluto andCharon system for the three NIKA2 array observations are given

11


Table 1. Summary of the principle characteristics and performance of the NIKA2 instrument.

Channel 260 GHz 150 GHz1.15 mm 2 mm

Arrays A1 A3 A1&3 A2Number of designed detectors 1140 1140 616

Number of valid detectors1 952 961 553Number of used detectors2 660 - 830 725 -780 430 - 480

FOV diameter [arcmin] 6.5 6.5 6.5 6.5FWHM [arcsec] 11.3 ± 0.2 11.2 ± 0.2 11.2 ± 0.1 17.7 ± 0.1

Beam efficiency3 [% ] 55 ± 5 53 ± 5 60 ± 6 75 ± 5rms calibration error [%] 4.5 6.6 5

Model absolute calibration uncertainty [%] 5RMS pointing error [arcsec] < 3

NEFD [mJy.s1/2] 4 33±2 8±1Mapping speed [arcmin2/h/mJy2] 5 67 - 78 1288 - 1440

Notes. (1) Number of detectors that are valid at least for two different beam map scans. (2) Number of detectors used in the scientific analysis afterstringent selection (Perotto et al. 2017, Ponthieu et al. 2017). (3) Ratio between the main beam power and the total beam power up to a radius of250′′ (4) Average NEFD during the February 2017 observation campaign, extrapolated at τ = 0. (5) Average mapping speed during the February2017 observation campaign, extrapolated at τ = 0.

Table 2. NIKA2 measured flux for a selection of sources. Statistical and calibration uncertainties are given.

Source Observing time [hours]a A1 Flux [mJy] A3 Flux [mJy] 260 GHz Flux [mJy] 150 GHz Flux [mJy]MWC349 3.44 1994±1.2±140 2048±1±143 2027±1±142 1389.5±0.2±97

SO1 (20h33′10.416′′, +40◦41′15′′ 3.44 78±1±7 29.7±0.3±2SO2 (20h33′11.928′′, +40◦41′44′′ 3.44 396±1±28 93.8±0.4±7

Pluto-Charon 1.44 15.8±1.6±1.1 15.4±1.2±1.1 15.5±1.0±1.1 4.8±0.2±0.3

Notes. (a) Observing time is on-source time excluding time for slewing, pointing, and focussing.

in Table 2. With an observation time of 1.44 hours we reach 1and 0.3 mJy (1-σ) at 260 GHz (1.15 mm) and 150 GHz (2 mm),respectively, over a 150 arcmin2 map.

These results illustrate the high sensitivity of NIKA2, withwhich mJy sources can be detected in less than one hour.

6. Conclusions and future plans

The NIKA2 instrument has been permanently installed at the30-meter IRAM telescope since September, 2015. A first tech-nical upgrade was achieved in September 2016. During this up-grade, we replaced the dichroic, changed the 150 GHz array andreplaced most of the smooth lenses with anti-reflecting-coatedones. A number of commissioning observational runs have beenachieved since the first light, that is, October, 2015. In the presentpaper we provided a general overview of the instrument, andshow the main results obtained during the commissioning cam-paigns.

The performance of the instrument, in terms of sensitiv-ity, surpasses the ambitious goals at 150 GHz, and is, even perbeam, better than previous generation instruments at 260 GHz.Building on this base, NIKA2 has been opened, in April 2017,to science verification observations, and in October 2017 to gen-eral science observations. We are preparing a first purely as-trophysical publication centred on high-quality mapping of thehigh-redshift (z = 0.58) galaxy cluster PSZ2 G144.83+25.11 viathe SZ effect (Ruppin et al. 2017). The instrument is offered,under IRAM responsibility, to a larger community. In paral-lel, we have entered a phase of commissioning of the polarisa-tion module of NIKA2 following the work performed in NIKA(Ritacco et al. 2017).

NIKA2, thanks to its versatile design and to the KID tech-nology adopted, will be upgraded during its lifetime. There are

several possible upgrades that we are considering: widening the260 GHz channels band in order to match the ”1 mm atmo-spheric window” in very good observing conditions, adding athird band, reducing the pixel size, adding a polarised channel at150 GHz, increasing the illumination of the primary mirror andothers.

Acknowledgements. We would like to thank the IRAM team in Spain for theirwork leading NIKA2 towards success; in particular Gregorio Galvez, PabloGarcia, Israel Hermelo, Dave John, Hans Ungerechts, Salvador Sanchez, PabloMellado, Miguel Munoz, Francesco Pierfederici, Juan Penalver. On top of that,we also would like to thank the remaining IRAM Granada staff for the outstand-ing support before, during and after the observations. In particular, we thank thetelescope operators and the logistics and administration groups. We acknowl-edge the crucial contributions of the technological groups in Neel, LPSC andIRAM Grenoble, and in particular: A. Barbier, E. Barria, D. Billon-Pierron, G.Bosson, J.-L. Bouly, J. Bouvier, G. Bres, P. Chantib, G. Donnier-Valentin, O.Exshaw, T. Gandit, G. Garde, C. Geraci, A. Gerardin, C. Guttin, C. Hoarau,M. Grollier, C. Li, J. Menu, J.L. Mocellin, E. Perbet, N. Ponchant, G. Pont, H.Rodenas, S. Roni, S. Roudier, J.P. Scordilis, O. Tissot, D. Tourres, C. Vescovi,A.J. Vialle. Their technical and scientific skills, as well as human qualities, rep-resent our main boost. The NIKA2 contracts have been administrated by thelaboratories involved. We acknowledge the contribution of P. Poirier, M. Berard,C. Bartoli, D. Magrez, and F. Vidale, among others. We enjoy frequent funda-mental physics discussions concerning superconducting devices with FlorenceLevy-Bertrand, Olivier Dupre, Thierry Klein, Olivier Buisson and other col-leagues at the Institut Neel. This work has been mainly funded by the ANR underthe contracts ”MKIDS”, ”NIKA”, ANR-15-CE31-0017 and LabEx ”FOCUS”ANR-11-LABX-0013. NIKA2 has benefited from the support of the EuropeanResearch Council Advanced Grant ORISTARS under the European Union’sSeventh Framework Programme (Grant Agreement no. 291294). We acknowl-edge funding from the ENIGMASS French LabEx, the CNES post-doctoralfellowship program, the CNES doctoral fellowship program and the FOCUSFrench LabEx doctoral fellowship program. GL, AB, HA and NP acknowledgefinancial support from the Programme National de Cosmologie and Galaxies(PNCG) funded by CNRS-CEA-CNES, from the ANR under the contract ANR-15-CE31-0017, the OCEVU Labex (ANR-11-LABX-0060) and the AMIDEXproject (ANR-11-IDEX-0001-02) funded by the Investissements d’Avenir pro-gram managed by the ANR.

12


References

Adam, R., Comis, B., Macıas-Perez, J.-F., et al. 2014, Astronomy &Astrophysics 569, id.A66

Andre, P., Menshchikov, A., Bontemps, S., et al. 2010, Astronomy &Astrophysics 518, id. L102

Bendo, G. J., Griffin, M. J., Bock, J. J., et al. 2013, Monthly Notices of the RoyalAstronomical Society 433, Issue 2, 3062

Bethermin, M., Wu, H.Y., Lagache, G., et al. 2017, Astronomy & Astrophysics,submitted, arxiv:1703.08795

Bourrion, O., Vescovi, C., Bouly, J.L., et al. 2012, Journal of Instrumentation 7,P07014, arXiv:1204.1415

Bourrion, O., Benoit, A., Bouly, J.L., et al. 2016, Journal of Instrumentation 11,P11001, arXiv:1602.01288

Bracco, A., Palmeirim, P., Andre, P., et al. 2017, Astronomy & Astrophysics, inpress, arxiv:1706.08407

Calvo, M., Giordano, C., Battiston, R., et al. 2010, Experimental Astronomy 28,Issue 2-3, 185

Calvo, M., Roesch, M., Desert, F.-X., et al. 2013, Astronomy & Astrophysics551, id.L12

Calvo, M., Benoit, A., Catalano, A., et al. 2016 Journal of Low TemperaturePhysics 184, Issue 3-4, 816

Carter, M., Lazareff, B., Maier, D., et al. 2012, Astronomy & Astrophysics 538,id.A89

Catalano, A., Calvo, M., Ponthieu, N., et al. 2014, Astronomy & Astrophysics569, id.A9

Chavez-Dagostino, M., Bertone, E., Cruz-Saenz de Miera, F., et al. 2016,Monthly Notices of the Royal Astronomical Society 462, Issue 3, 2285

Day, P. K., LeDuc, H. G., Mazin, B. A., Vayonakis, A., & Zmuidzinas, J. 2003,Nature, 425, 817

Dempsey, J. T., Friberg, P., Jenness, T., et al. 2013, Monthly Notices of the RoyalAstronomical Society 430, Issue 4, 2534

Dicker, S.R., Ade, P.A.R., Aguirre, J., et al. 2014, Proceedings of the SPIE 9153,id. 91530J

Doyle, S., Mauskopf, P., Zhang, J., et al. 2010 Proceedings of the SPIE 7741, id.77410M

Durand, T., 2008, PhD Thesis, Universite de GrenobleGeach, J.E, Dunlop, J.S, Halpern, M., et al. 2017, Monthly Notices of the Royal

Astronomical Society 465, Issue 2, 1789Goupy, J., Adane, A., Benoit, A., et al. 2016, Journal of Low Temperature

Physics 184, Issue 3-4, 661Greve, A., Kramer, C., & Wild, W. 1998, A&AS, 133, 271Hasselfield, M., Hilton, M., Marriage, T.A., et al. 2013, Journal of Cosmology

and Astroparticle Physics, 7, 8Holland, W.S., Bintley, D., Chaplin, E.L., et al. 2013, Monthly Notices of the

Royal Astronomical Society 430, Issue 4, 2513Konyves, V., Andre, P., Menshchikov, A., et al. 2015, Astronomy & Astrophysics

584, id. A91Kramer, C., Penalver, J., & Greve, A. 2013, Internal IRAM DocumentKryukova, E., Megeath, S., Hora, J., et al., 2014, AJ, 148, 11KMonfardini, A., Swenson, L.J., Bideaud, A., et al. 2010, Astronomy &

Astrophysics 521, id.A29Monfardini, A., Benoit, A., Bideaud, A., et al. 2011, The Astrophysical Journal

Supplement 194, Issue 2, id. 24Moreno, R., 1998, PhD thesis, Universite de ParisMoreno, R., 2010, Neptune and Uranus planetary brightness temperature tabula-

tion. Tech. rep., ESA Herschel Science CenterPardo, J. R., Cernicharo, J. & Serabyn, E., 2002, IEEE Transactions on Antennas

and Propagation 49, 1683Perotto, L., et al., Astronomy & Astrophysics, in preparationPlanck Collaboration, 2016, A&A, 594, A28Planck Collaboration, 2016, A&A, 594, A24Planck Collaboration, 2016, A&A, 594, A22Ponthieu, N., et al., Astronomy & Astrophysics, in preparationPisano, G., Savini, G., Ade, P.A.R., et al. 2008, Applied Optics 47, issue 33,

6251Reichardt, C. L., Stalder, B., Bleem, L. E., et al. 2013, The Astrophysical Journal

763, 127Reveret, V., Andre, P., Le Pennec, J., et al. 2014, Proceedings of the SPIE 9153,

id. 915305Ritacco, A., Ponthieu, N., Catalano, A., et al. 2017, Astronomy & Astrophysics

599, id.A34Roesch, M., Benoit, A., Bideaud, A., et al., 2012, ISSTT2011 Workshop,

arXiv:1212.4585Rosolowsky, E., et al 2010, ApJS, 188, 123RRuppin, F., et al. 2017, In preparation, to be submitted to Astronomy &

Astrophysics

Schuster, K.-F., Boucher, C., Brunswig, W., et al. 2004 Astronomy &Astrophysics 423, 1171

Siringo, G., Kreysa, E., Kovacs, A., et al. 2009, Astronomy & Astrophysics 497,Issue 3, 945

Staguhn, J. G., Benford, D. J., Dowell, C. D., et al. 2016, Journal of LowTemperature Physics 184, Issue 3-4, 811

Swenson, L. J., Cruciani, A., Benoit, A., et al. 2010, Applied Physics Letters 96,Issue 26, id. 263511

13


Fig. 14. From top to bottom, detectors positions for arrays A1(260 GHz-H), A3 (260 GHz-V), and A2 (150 GHz). The threeplots show the detectors that have seen the sky and passed thequality criteria for at least two focal plane reconstructions dur-ing Run10: 952, 961, and 553 for A1, A3 and A2, respectively.The outer dashed line circle corresponds to the nominal FoV of6.5 arc-minutes.

Fig. 15. (Colour online) From top to bottom, main beam FWHMdistribution of all valid KID detectors of arrays A1, A3, andA2. The main beam FWHM is the geometrical combination ofthe two-orthogonal FWHM estimates obtained from an ellipticalGaussian fit on side-lobe masked individual maps per KID (seetext). The red curves show a Gaussian fit to the histogram data.

14


Fig. 16. (Colour online) Measured beam pattern. From upper leftto lower right, beam maps of arrays A1 and A3, the combinationof the 260 GHz arrays (A1&A3) and the 150 GHz array (A2) areshown in decibel. These 10′ × 10′ maps, have been constructedfrom the normalised combination of four relatively long scansof bright point sources. Details on the structures present in themaps are given in the text.

Fig. 17. (Colour online) Comparison of measured and refer-ence flux densities of the primary calibrators Uranus (red) andNeptune (blue). Their ratios are shown for the three arrays A1(260 GHz-H), A2 (150 GHz), A3 (260 GHz-V). The mean ratioµ and relative scatter are provided for each array. The referenceflux densities are from Moreno, 2010, Bendo et al. (2013). Thescan numbers are time-ordered: 1 to 10 refer the period 23-28February, 2017, (fair weather) and 11 to 18 to the period 19-25April, 2017, (mediocre weather).

Fig. 18. (Colour online) From top to bottom power spectra of theNIKA2 time ordered data before (black) and after (blue) subtrac-tion of atmospheric fluctuations, which show-up at frequenciesbelow 1 Hz.

15


(

MWC349 1.15 mm

20h33m15s 00s 32m45s 32m30s 32m15s

40◦45’00”

42’00”

39’00”

36’00”

33’00”

Right Ascension (J2000) [hr]

Dec

linat

ion

(J20

00)[

degr

ee]

−0.050

−0.025

0.000

0.025

0.050

0.075

0.100

0.125

0.150

0.175

0.200

Sur

face

brig

htne

ss[J

y/be

am]

MWC349 2 mm

20h33m15s 00s 32m45s 32m30s 32m15s

40◦45’00”

42’00”

39’00”

36’00”

33’00”


Dec

linat

ion

(J20

00)[

degr

ee]

−0.03

−0.02

−0.01

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Sur

face

brig

htne

ss[J

y/be

am]

Fig. 19. (Colour online) Maps at 260 GHz (top) and 150 GHz(bottom) centered in MWC349. Details on the observed sourcesare given in the text. The contours in these maps indicate signal-to-noise ratios of 5, 7 and 10. The FWHM are given in the lowerleft corners.

(

Pluto 1.15 mm

19h19m55s 50s 45s 40s

-21◦12’00”

13’00”

14’00”

15’00”

16’00”


Dec

linat

ion

(J20

00)[

degr

ee]

−4.5

−3.0

−1.5

0.0

1.5

3.0

4.5

6.0

Sur

face

brig

htne

ss[m

Jy/b

eam

]

Pluto 2 mm

19h19m55s 50s 45s 40s

-21◦12’00”

13’00”

14’00”

15’00”

16’00”


Dec

linat

ion

(J20

00)[

degr

ee]

−1.5

−1.2

−0.9

−0.6

−0.3

0.0

0.3

0.6

0.9

1.2

1.5

Sur

face

brig

htne

ss[m

Jy/b

eam

]

Fig. 20. (Colour online) Maps at 260 GHz (top) and 150 GHz(bottom) of the Pluto and Charon planetary system. The contoursin these maps indicate signal-to-noise ratios of 5, 7 and 10. TheFWHM are given in the lower left corners.

16

Documents

The NIKA2 large-ﬁeld-of-view millimetre continuum camera ...NIKA2 and the IRAM 30-meter telescope represent today ideal tools to address these scientiﬁc questions, and many oth-1